Industrial Chemical Process Design 9780071376204, 0071376208

Citation preview

Chapter

1 The Database

Nomenclature µ

viscosity, cP

°API

gravity standard at 60°F

cP

viscosity, centipoise

cSt

viscosity, centistokes

D

density, lb/ft3

exp

constant for exponential powers of e base value, 2.7183

MW

molecular weight

P

system pressure, psia

P

pressure, psia

PC

critical pressure, psia

PR

reduced pressure (P/PC), psia

°R

T + 460°F

SG 60/60

specific gravity referenced to pure water at 60°F

T

temperature, °F

TB

true boiling point, °F, of ASTM curve cut component or pure component

TC

critical temperature, °F or °R

TR

reduced temperature (T/TC), °R

V

molar volume, ft3/lb-mol

Z

gas compressibility factor

To meet this introductory challenge, we must first establish a database from which to launch our campaign. In doing so, consider the physical properties of liquids, gases, chemicals, and petroleum gener-

1

2

Chapter One

ally in making this application: viscosity, density, critical temperature, critical pressure, molecular weight, boiling point, acentric factor, and enthalpy. The great majority of the process engineer’s work is strictly with organic chemicals. This book is therefore directed toward this database of hydrocarbons (HCs). Only eight physical properties are presented here. Aren’t there many others? The answer is yes, but remember, this book is strictly directed toward that which is indeed practical. Many more properties can be listed, such as critical volume or surface tension. Our quest is to take these more practical types (the eight) as our database and thereby successfully achieve our goal, practical process engineering (PPE). At this point, it is important to present a disclaimer. Many notable engineers could claim the author is loco to think he can resolve all database needs with only the eight physical properties given or otherwise derived in this book. Let me quickly state that many other extended database resources are indeed referenced in this book for the user to pursue. Only in such retrieval of these and many other database resources, such as surface tension and solubility parameters, can PPE be applied. An example is that surface tension and solubility parameters must both be determined before the liquid/liquid software program given herein can be applied. This liquid-liquid extraction program (Chemcalc 16 [1]) is included as part of the PPE presentation. (See Chap. 7.) It is therefore important to keep in mind that many database references are so pointed to in this book—Perry’s, Maxwell’s, and the American Petroleum Institute (API) data book, to name a few. Again, why then present only these eight physical properties for our concern? The answer is that we can perform almost every PPE scenario by applying these eight physical properties, which are in most every data source and are readily available. Furthermore, an exhaustive listing would be a much greater book than the one you are reading, such as Lange’s Handbook of Chemistry and Physics [2]. Incidentally, Lange’s is a very good reference book which I highly endorse. Viscosity Liquid viscosity

The first of these properties is viscosity. All principal companies use mainly one of two viscosity units, centipoise (cP) or centistokes (cSt). Centipoise is the more popular. If your database presents only one, say cP, then you may quickly convert it to the other, cSt, by a simple equation: cP cSt = ᎏ sp gr

The Database

3

Specific gravity (sp gr) is simply the density referenced to water, sp gr of water being 1.0 at 60°F. This means that to get the specific gravity of a liquid, simply divide the density of the liquid at any subject temperature of the liquid by the density of water referenced to 60°F. With sp gr so defined, we can subsequently convert cP to cSt or cSt to cP by this simple equation. Thus the conversion is referenced to a temperature. Viscosity of any liquid is very dependent and varies with the slightest variance of the liquid temperature. Viscosity has been defined as the readiness of a fluid to flow when it is acted upon by an external force. The absolute viscosity, or centipoise, of a fluid is a measure of its resistance to internal deformation or shear. A classic example is molasses, a highly viscous fluid. Water is comparatively much less viscous. Gases are considerably less viscous than water. For the viscosity of most any HC, see Fig. A-3 in Crane Technical Paper No. 410 [3]. If your particular liquid is not given in this viscosity chart and you have only one viscosity reading, then locate this point and draw a curve of cP vs. temperature, °F, parallel to the other curves. This is a very useful technique. I have found it to be the more reliable, even when compared to today’s most expensive process simulation program. Furthermore, I find it to be a valuable check of suspected errors in laboratory viscosity tests. If you don’t have the Crane tech paper (available in any technical book store), then get one. You need it. I have found that most every process engineer I have met in my journeys to the four corners of the earth has one on their bookshelf, and it always looks very used. The following equations compose a good quick method that I find reasonably close for most hydrocarbons for API gravity basis. Note that the term API refers to the American Petroleum Institute gravity method [4]. These viscosity equations are derived using numerous actual sample points. These samples ranged from 10 to 40° API crude oils and products. I find the following equations, Eqs. (1.1) to (1.4), to be in agreement with Sec. 9 of Maxwell’s Data Book on Hydrocarbons [5]. How to determine any HC liquid viscosity.

Viscosity, cP, for 10°API oil: µ = exp (18.919 − 0.1322T + 2.431e-04 T 2)

(1.1)

Viscosity, cP, for 20°API oil: µ = exp (9.21 − 0.0469T + 3.167e-05 T 2)

(1.2)

Viscosity, cP, for 30°API oil: µ = exp (5.804 − 0.02983T + 1.2485e-05 T 2)

(1.3)

4

Chapter One

Viscosity, cP, for 40°API oil: µ = exp (3.518 − 0.01591T − 1.734e-05 T 2) where

(1.4)

µ = viscosity, cP T = temperature, °F exp = constant of natural log base, 2.7183, which is raised to the power in parentheses

You can interpolate linearly for any API oil value between these equations and with extrapolation outside to 90°API. Temperature coverage is good from 50 to 300°F. If outside of this range, use the American Society for Testing and Materials (ASTM) Standard Viscosity-Temperature Charts for Liquid Petroleum Products (ASTM D-341 [6]). The values derived by Eqs. (1.1) to (1.4) are found to be within a small percentage of error by the ASTM D-341 method. It is good practice to always obtain at least one lab viscosity reading. With this reading, draw a relative parallel curve to the curve family in ASTM D-341. The popular Crane Technical Paper No. 410 reproduces this ASTM chart as Fig. B-6. If two viscosity points with associated temperature are known, then use the Crane log plot figure, also an API given method (ASTM D-341), to determine most any liquid hydrocarbon viscosity. Gas viscosity How to determine any HC gas viscosity. For most any HC gas viscosity,

use Fig. 1.1 (Fig. A-5 in Crane Technical Paper No. 410). The constant Sg

Figure 1.1 Hydrocarbon gas viscosity. (Adapted from Crane Technical Paper No. 410, Fig. A-5. Reproduced by courtesy of the Crane Company [3].)

The Database

5

is simply the molecular weight (MW) of the gas divided by the MW of air, 29. Note carefully, however, that Fig. 1.1 is strictly limited to atmospheric pressure. The gas atmospheric reading from this figure, or from other resources such as the API Technical Data Book [7], is deemed reasonably accurate for pressures up to, say, 400 psig. In addition to the API Technical Data Book and Gas Processors Suppliers Association (GPSA) methods [8], a new gas viscosity method is presented herein that may be used for a computer program application or a hand-calculation method: For 15 MW gas: µ = 0.0112 + 1.8e-05 T

(1.5)

µ = 0.00923 + 1.767e-05 T

(1.6)

µ = 0.00773 + 1.467e-05 T

(1.7)

µ = 0.0057 + 0.00001 T

(1.8)

For 25 MW gas:

For 50 MW gas:

For 100 MW gas:

where

µ = viscosity, cP T = temperature, °F

Equations (1.5) to (1.8) are good from vacuum up to 500 psia pressure and temperatures of −100 to 1000°F. Pressures at or above 500 psia should have corrections added from Eqs. (1.9), (1.10), or (1.11). Eqs. (1.5) to (1.8) are reasonably accurate to within 3% of the API data and are good for a pressure range from atmospheric to approximately 400 psig. You may make linear interpolations between temperaturecalculated points for reasonably accurate gas viscosity readings at atmospheric pressures. Many will say (even notable process engineers, regrettably) that higher pressures (above 400 psig) will have little effect on the gas viscosity, and that although the viscosity does change, the change is not significant. Trouble here! In many unit operations, such as high-pressure (≤500 psig) separators and fractionators, the gas viscosity variance with pressure is most critical. I have found this gas viscosity variance to be significant in crude oil–production gas separators, even as low as 300 psig. You may make corrections with the following additional equations. These corrections are to be added to the atmo-

The Database 6

Chapter One

spheric gas viscosity reading in Fig. 1.1 or the gas viscosity Eqs. (1.5) to (1.8) [8]. Add the following calculated viscosity correction to Fig. 1.1 or Eqs. (1.5) to (1.8): Gas viscosity correction for 100°F system: µc = −1.8333e-05 + 1.2217e-06 P + 1.737e-09 P2 − 2.1407e-13 P3

(1.9)

Gas viscosity correction for 400°F system: µc = −1.281e-05 + 1.5484e-06 P + 2.249e-10 P2 − 6.097e-14 P3

(1.10)

Gas viscosity correction for 800°F system: µc = −1.6993e-05 + 1.1596e-06 P + 2.513e-10 P2

(1.11)

where µc = viscosity increment, cP, to be added to Fig. 1.1 values or to Eqs. (1.5) to (1.8) P = system pressure, psia You may again interpolate between equation viscosity values for inbetween pressures. Whenever possible, and for critical design issues, these variables should be supported by actual laboratory data findings. In applying these equations and Fig. 1.1, please note that you are applying a proven method that has been used over several decades as reliable data. Henceforth, whenever you need to know a gas viscosity, you’ll know how to derive it by simply applying this method. You may also use these equations in a computer for easy and quick reference. See Chap. 9 for computer programming in Visual Basic. Applying programs such as these is simple and gives reliable, quick answers. Now somebody may say here, why should a high-pressure (450 psig and greater) gas viscosity be so important, and, by all means, what is practical about finding such data? Well, this certainly deserves an answer, so please see Chap. 4, page 153, for the method of calculating gas and liquid vessel diameters. Note that Eq. (4.3) in the Vessize.bas program has an equation divided by the gas viscosity. A change of only 10% in the gas viscosity value greatly changes the vessel’s required diameter, as may be seen simply by running this vessel-sizing program. Considerable emphasis is therefore placed on these database calculations. They do count. Take my suggestion that you prepare for a good understanding of this database and how to get it.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

7

Density Liquid density

Liquid density for most HCs may be found in Fig. 1.2. This chart is a general reference and may be used for general applications that are not critical for discrete defined components. In short, if you don’t have a better way of getting liquid density, you can get it from Fig. 1.2. Note that you need to have a standard reference of API gravity reading to predict the HC liquid density at any temperature. Generally, you should have such a reading given as the API gravity at 60°F on the crude oil assay or the petroleum product cut lab analysis. If you don’t have any of these basic items, you must have something on which to base your component data, such as a pure component analysis of the mixture. If not, then please review the basis of your given data, as it is most evident that you are missing critical data that must be made available by obtaining new lab analysis or new data containing API gravities. It is important here to briefly discuss specific gravity and API gravity of liquids. First, the API gravity is always referenced to one temperature, 60°F, and to water, which has a density of 62.4 lb/ft3 at this temperature. Any API reading of a HC is therefore always referenced to 60°F temperature and to water at 60°F. This gravity is always noted as SG 60/60, meaning it is the value interchangeable with the referenced HC’s API value per the following equations:

Specific gravity of petroleum fractions. (Plotted from data in J. B. Maxwell, “Crude Oil Density Curves,” Data Book on Hydrocarbons, D. Van Nostrand, Princeton, NJ, 1957, pp. 136–154.)

Figure 1.2

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 8

Chapter One

141.5 °API = ᎏᎏ − 131.5 SG 60/60

(1.12)

141.5 SG 60/60 = ᎏᎏ 131.5 + °API

(1.13)

Thus, having the API value given, we may find the subject HC gravity at any temperature by applying Fig. 1.2. Keep in mind that liquid gravities are always calculated by dividing the known density of the liquid at a certain temperature by water at 60°F or 62.4 lb/ft3. I also find the following equation to be a help (again, in general) in deriving a liquid density. Liquid density estimation

MW D = ᎏᎏᎏᎏᎏᎏ (10.731 ∗ TC/PC) ∗ 0.260^[1.0 + (1.0 − TR)^0.2857] where

(1.14)

D = liquid density, lb/ft3 MW = molecular weight TC = critical temperature, degree Rankine (°R) PC = critical pressure, psia TR = reduced temperature ratio = T/TC T = system temperature, °R, below the critical point

Let’s now run a check calculation to see how accurate this equation is. n-Octane

MW = 114.23 PC = 360.6 psia TC = 1024°R SG 60/60 = 0.707 Trial at 240∞F, Liquid n-Octane

From Eq. (1.14): MW D = ᎏᎏᎏᎏᎏᎏ (10.731 ∗ TC/PC) ∗ 0.260^[1.0 + (1.0 − TR)^0.2857] D = 38.00 lb/ft3 From Maxwell, page 140 [9]: or

n-octane density at 240°F = 0.625 SG 0.625 ∗ 62.4 = 39.00 lb/ft3

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

9

From Fig. 1.2: 141.5 °API @ 0.707 SG 60/60 = ᎏ − 131.5 = 68.64°API 0.707 At this API curve, read 0.615 gravity (horizontal line from intersect point) at 240°F, or 0.615 ∗ 62.4 = 38.38 lb/ft3 Summary of Eq. (1.14) Check

39.0 − 38.0 Deviation from Maxwell [9] = ᎏᎏ ∗ 100 = 2.5% error 39 38.38 − 38.00 Deviation from Nelson [10] = ᎏᎏ ∗ 100 = 1.0% error 38.38 From the preceding check of n-octane liquid density, we have established that Eq. (1.14) is a reasonable source for calculating n-octane liquid density. Both Nelson and Maxwell data points could also have as much error, 1 to 3%. The conclusion therefore is that Eq. (1.14) is a reasonable and reliable method for liquid density calculations. You may desire to investigate other known liquid densities having the same known variables, TC, PC, and MW. You are encouraged to do so. Gas density

While the density of any liquid is easily derived and calculated, the same is not true for gas. Gas, unlike liquid, is a compressible substance and varies greatly with pressure as well as temperature. At low pressures, say below 50 psia, and at low temperature, say below 100°F, the ideal gas equation of state holds true as the following equation: MW ∗ P D = ᎏᎏ 10.73 ∗ (460 + T) where

(1.15)

D = gas density, lb/ft3 MW = gas molecular weight P = system pressure, psia T = °F

For this low-temperature and -pressure range, any gas density may quickly be calculated. Error here is less than 3% in every case checked. What about higher temperatures and pressures? Aren’t these higher values where all concerns rest? Yes, most all process unit operations, such as fractionation, separation, absorption stripping, chemical reaction, and heat exchange generate and apply these higher-temperature Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 10

Chapter One

and -pressure conditions. Then how do we manage this deviation from the ideal gas equation? The answer is to insert the gas compressibility factor Z. Add gas compressibility Z to Eq. (1.15): MW ∗ P D = ᎏᎏᎏ Z ∗ [10.73 ∗ (460 + T)]

(1.16)

where Z = gas compressibility factor The question now is how do you derive, calculate, or find the correct Z factor at any temperature and pressure? The first answer is, of course, get yourself a good, commercially proven, process-simulation software program. As these programs cost too much, however, for anyone who works for a living, you must seek other resources. This is a core reason why this book has been written. Look at the practical side. After all, who has $25,000 pocket change to throw out for such candor? It is therefore my sincere pleasure to present to you, as the recipient of the software accompanying this book, the following two computer programs. Z.mak. This is a program derived from data established in the API

Technical Data Book, procedure 6B1.1 [11]. Please note that Z.mak, although similar, is an independent and separate program from this API procedure. A program listing as in the Z.mak executable file is shown in Table 1.1. Inside the phase envelope, the compressibility factor calculated in Z.mak is more accurate than that calculated in RK.mak (the Redlich-Kwong equation of state). RK.mak is given and discussed later. The Z.mak program may be used with reasonable accuracy, as can the API procedure 6B1.1. Z.mak accuracies range from 1 to 3% error. Most case accuracies are 1% error or less. One caution, however, is necessary, and this is regarding Z values in or near the critical region of the phase envelope. Important Note: Use Z.mak when at less than the critical pressure and/or in the phase envelope. The acentric factor is also calculated from the input TC, PC, and boiling point. The acentric factor is used in the Z factor derivation. See line 110 in Table 1.1. Please note that the Z factor so calculated here is to be applied in Eq. (1.16) for calculating the gas density. The Z factor for butene-1 is now calculated in the actual computer screen display of the Z.mak computer program. (See Fig. 1.3.) RK.mak. When out of the phase envelope, use this program, the well-

known Soave-Redlich-Kwong (SRK) equation-of-state simplified program [12]. The student here may immediately detect the standard SRK Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

TABLE 1.1

11

Z.mak Program Code Listing

Sub Command1_Click () 10 'Program for calculating gas compressibility factor, Z 15 'For a Liquid - Vapor Equilibrium Saturation Condition, gas Z If TxtTC = "" Then TxtTC = 295.6 If TxtPC = "" Then TxtPC = 583 If TxtTB = "" Then TxtTB = 20.7 If TxtP = "" Then TxtP = 200 20 'Data Input lines 30 through 50 TC = TxtTC: PC = TxtPC: TB = TxtTB: P = TxtP 30 TC = TC + 460 'Critical T in deg R 40 TB = TB + 460 'Atmospheric boiling Temperature in deg R 45 TR = TB / TC 50 PR = P / PC: PR2 = 14.7 / PC 'System P and Reduced PR Calc psia 60 'Calculate Acentric Factor, ACENT PR0 = 6.629 - 11.271 * TR + 4.65 * TR ^ 2: PR1 = 16.5436 - 46.251 * TR + 45.207 * (TR ^ 2) - 15.5 * (TR ^ 3) ACENT = (((Log(PR2)) / 2.3026) - (-PR0)) / (-PR1) 70 'ACENT = .42857 * (((.43429 * Log(PC)) - 1.16732) / ((TC / TB) - 1#)) - 1# 80 'Equations for Z calc follow 90 Z0 = .91258 - .15305 * PR - (1.581877 * (PR ^ 2)) + (2.73536 * (PR ^ 3)) - (1.56814 * (PR ^ 4)) 100 Z1 = -.000728839 + .00228823 * PR + (.217652 * (PR ^ 2)) + (.0181701 * (PR ^ 3)) - (.1544176 * (PR ^ 4)) 110 Z = Z0 - ACENT * Z1 120 'Print ACENT, Z TxtACENT = Format(ACENT, "##.######"): TxtZ = Format(Z, "##.######") End Sub SOURCE: Method from data in Calculation method GB1.1, “API Density,” American Petroleum Institute, Technical Data Book, API Refining Department, Washington, DC, 1976.

equation of state on line 110 and the derivation of coefficients A and B on line 100 in Table 1.2. This program solves a cubic equation by first assuming a value for V, line 80, and then iterating a pressure calculation of P on line 110 until the calculated DELV of V deviation is less than 0.0001. Thus, this is a unique way to calculate V and the density thereof per Eq. (1.16). Please note that both Z.mak and RK.mak exhibit the same problem for finding the gas density of propane at 100 psia and 200°F. Note also that Z calculations from each are appreciably different, 0.86 vs. 0.94 (see Figs. 1.3 and 1.4). Why the difference? Remember the previous warning about using the Z.mak program out of the phase envelope? Well, this is a classic example, as these conditions are definitely out of Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 12

Chapter One

Figure 1.3

Z.mak screen.

propane’s phase envelope. Therefore, RK.mak should be correct here, and it is indeed correct. You may verify the answer with any propane pressure-temperature-enthalpy chart. A fluid density of 0.66 lb/ft3 and a Z factor of 0.94 are correct. For those of you who are scavengers and are rapidly scanning this book to claim whatever treasure you may find, may I say my good cheers and gung ho (a World War II saying for “go get ’em!”). For those of you who are weeding out every word in careful analysis of what I’m trying to deliver in this book, however, I must share the following thoughts. I have been a full-time employee in three major engineering, procurement, and construction (EPC) firms, and in each one I had very limited access to these high-priced simulation programs that do almost every calculation imaginable and a few more on top of that. As a practicing process design engineer, I can remember more times I needed these simulation programs on my computer and didn’t have one, than I can remember having one when I needed it. Seems these companies always have a young engineer who is indeed a whiz on these simulation programs, the one and only person who runs the simulation program. You need a data set run? Well, you must give the engineer your data in elite form and then wait in queue for the output answers. Uh-oh, you’ve now got the answers and suddenly realize you didn’t cover the entire range critically needed? Do it all over again and wait in queue for your answers, hoping

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

TABLE 1.2

13

RK.mak Program Code Listing

Sub Command1_Click () 10 'RK EQUATION OF STATE PROGRAM FOR GAS DENS CALC 20 'Print " RK EQUATION OF STATE PROGRAM FOR GAS DENS CALC": Print : Print 30 'INPUT " P PSIA, T DEG F ",P,T 40 'INPUT " PC PSIA, TC DEG F ",PC,TC T = TxtT: P = TxtP: TC = TxtTC: PC = TxtPC: MW = TxtMW 50 T = T + 460: TC = TC + 460 60 ' INPUT " MW OF MIXTURE ",MW 70 'FIRST TRIAL GUESS FOR V, CF PER lb MOL DENSITY: 80 V = 10.73 * T * .001 / P 90 Rem A & B CONST CALC 100 B = .0867 * 10.73 * TC / PC: A = 4.934 * B * 10.73 * (TC ^ 1.5) 110 PCA = ((10.73 * T) / (V - B)) - (A / ((T ^ .5) * V * (V + B))) 120 DELP = PCA - P 130 DPDV = (((B + 2 * V) * A / (T ^ .5)) / ((V * (V + B)) ^ 2)) - (10.73 * T / ((V - B) ^ 2)) 140 V = V - DELP / DPDV 160 DELV = (DELP / DPDV) / V 170 If Abs(DELV) > .0001 GoTo 110 180 'PRINT:PRINT USING " FLUID DENS, lb/CF = ####.###### "; MW/V TxtDen = Format(MW / V, "#####.#####") 190 Z = (P * MW) / (10.731 * T * (MW / V)) TxtZ = Format(Z, "##.####") 200 'PRINT: PRINT USING " Gas Compressibility Factor, Z = ##.##### "; Z 210 'End End Sub SOURCE:

Method from O. Redlich and J. N. S. Kwong, Chem. Rev. 44:233, 1949.

you got it this time. You have now come to my hit line, use this book and the software herein to derive your needs! The previous example of critically needed density for, say, a hydraulic line sizing or heat exchanger problem is well in order with our modern-day, most advanced, high-priced computer programs. An added thought here is that most medium-sized EPC companies have only one or two keys to run these large computer software programs. Therefore, this book and the accompanying software will help you expedite much of the work independent of these large, costly programs. Just think, you’ve got your own personal key in this book and software! This book also is a good supplement to these complete and comprehensive simulation programs. As an added plus for you, the major solutions to your problems are given in the CD supplied with this book.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 14

Chapter One

Figure 1.4

RK.mak screen.

Industrial Chemical Process Design is indeed a toolkit offering the user practical process engineering. Having covered the difficulties of deriving an accurate gas density in Eqs. (1.15) and (1.16), it is important here to understand the practical application of same. First, for hydraulic line sizing, when the pressure of the line is 400 psig or less, consider using a conservative Z factor of 0.95 or 1.0. Look at Eq. (1.16). When Z decreases, the gas density increases, and thus the line size decreases. A conservative approach would be to use a larger Z than calculated or assume Z = 1.0 for a safe and conservative design. In most cases no line size increase results, while in some cases only one line size increase is the outcome. I suggest this is good practice. I have designed many flare systems and performed numerous emergency relief valve sizing calculations applying this Z = 1.0 criterion. Herein I suggest you also consider using Z = 1.0 for all relief valve and flare line sizing. This is a conservative and safe assumption. In practice, I have found every operating company to admire the assumption even to the point of endorsing it fully.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

15

Critical Temperature, TC To this point we have applied the critical temperature to both viscosity and density calculations. Already this critical property TC is seen as valued data to have for any hydrocarbon discrete single component or a mixture of components. It is therefore important to secure critical temperature data resources as much as practical. I find that a simple table listing these critical properties of discrete components is a valued data resource and should be made available to all. I therefore include Table 1.3 listing these critical component properties for 21 of our more common components. A good estimate can be made for most other components by relating them to the family types listed in Table 1.3. Also included here is an equation for calculating TC, °R, using SG 60/60 and the boiling point of the unknown HC. The constants A, B, and C are given for paraffins and aromatic-type families. For naphthene, olefin, and other family-type HCs A, B, and C constants, the process engineer is referred to the API Technical Data Book, Chap. 4, Method 4A1.1 [13]. TC = 10^[A + B ∗ log (sp gr) + C ∗ log TB]

TABLE 1.3

(1.17)

Critical Component Properties MW

TB, °F

Sp gr

PC, psia

TC, °F

Acentric fraction

Methane Ethane Propane n-Butane Isobutane

16.04 30.07 44.10 58.12 58.12

−258.69 −127.48 −43.67 31.10 10.90

0.3 0.3564 0.5077 0.5844 0.5631

667.8 707.8 616.3 550.7 529.1

−116.63 90.09 206.01 305.65 274.98

0.0104 0.0986 0.1524 0.2010 0.1848

n-Pentane Isopentane Neopentane

72.15 72.15 72.15

96.92 82.12 49.10

0.6310 0.6247 0.5967

488.60 490.40 464.00

385.70 369.10 321.13

0.2539 0.2223 0.1969

n-Hexane 2-Methylpentane 3-Methylpentane Neohexane 2,3-Dimethylbutane

86.17 86.17 86.17 86.17 86.17

155.72 140.47 145.89 121.52 136.36

0.6640 0.6579 0.6689 0.6540 0.6664

436.90 436.60 453.10 446.80 453.50

453.70 435.83 448.30 420.13 440.29

0.3007 0.2825 0.2741 0.2369 0.2495

209.17 194.09 197.32 200.25 174.54 176.89 186.91 177.58

0.6882 0.6830 0.6917 0.7028 0.6782 0.6773 0.6976 0.6949

396.80 396.50 408.10 419.30 412.20 396.90 427.20 428.40

512.80 495.00 503.78 513.48 477.23 475.95 505.85 496.44

0.3498 0.3336 0.3257 0.3095 0.2998 0.3048 0.2840 0.2568

Component

n-Heptane 2-Methylhexane 3-Methylhexane 3-Ethylpentane 2,2-Dimethylpentane 2,4-Dimethylpentane 3,3-Dimethylpentane Triptane

100.2 100.2 100.2 100.2 100.2 100.2 100.2 100.2

SOURCE: Data from Table 1C1.1, American Petroleum Institute, Technical Data Book, API Refining Department, Washington, DC, 1976.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 16

Chapter One

where sp gr = SG 60/60 TB = normal boiling temperature, °R Note: Log notation is base 10. Constants

Paraffins Aromatics Olefins

A

B

C

1.47115 1.14144 1.18325

0.43684 0.22732 0.27749

0.56224 0.66929 0.65563

This equation is acclaimed as good for paraffins up to 21 carbon atoms molecularly, and up to 15 carbon atoms for all others. Now for the best resource of all. Table 1.3 has many applications. In recent years new plant designers and plant upgrade designers have chosen many of the components shown in Table 1.3 to represent group compounds meeting the same family criteria. Each grouping may have hundreds of discrete identifiable compounds; however, only one is used to represent the entire group. Such grouping is being found to be acceptable error and most certainly is much better than a rough estimate. Critical Pressure, PC Table 1.3 is also an excellent source for critical pressure PC. If the particular HC compound or mixture is not listed in this table, consider relating it to a similar compound in Table 1.3. If molecular weight and the boiling points are known, you may find a close resemblance in Table 1.3. Also consider the API Technical Data Book, which lists thousands of HC compounds. Grouping as one component per se would also be feasible from Procedure 4A2.1 of the API book. Herein, components grouped together as a type of family could be represented as one component of the mixture. This one representing component may be called a pseudocomponent. Several of these pseudocomponents added together would make up the 100% molar sum of the mixture. As with TC, I also present herein a method to calculate PC applying the molecular group method [6]. 14.7 ∗ MW PC = ᎏᎏᎏ2 [(sum DELTPI) + 0.34] where

(1.18)

PC = critical pressure, psia MW = molecular weight DELTPI = compound molecular group structure contribution

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

17

Group Contributions DELTPI Non-ring-increment group contributions ᎏCH2

Ring-increment group contributions

0.227

ᎏCH2ᎏ

0.184

ᎏCH

0.192

ᎏCH2

0.227

苷CH2

0.198

苷CH

0.198

苷CH

0.154

苷CH

0.153

苷CH苷

0.154

Note: “sum” notation indicates the sum of DELTPI for each group contribution. Additional molecular group contributions can be found in Reid, Prausnitz, and Sherwood [14]. An example of group contributions is now run for benzene: sp gr = 0.8844 TB = 176.2°F + 460 = 636.2°R MW = 78.11 Data taken from Ref. 14 TC = 10^[A + B ∗ log (sp gr) + C ∗ log TB]

(1.17)

Aromatics A = 1.14144, B = 0.22732, C = 0.66929 TC = 1013°R or 553°F

From Table 1.3, TC = 552°F

14.7 ∗ MW PC = ᎏᎏᎏ2 [(sum DELTPI) + 0.34]

Checks okay (1.18)

Benzene has 6 苷CH groups at 0.154 each, or 0.924 total, and sum DELTPI = 0.924 PC = 719 psia

From Ref. 14, PC = 710 psia

Checks okay

This example of benzene shows that given the specific gravity at 60/60, the normal atmospheric boiling temperature, and the substance molecular weight, then the TC and PC critical properties can be calculated. These exhibited equations, Eqs. (1.17) and (1.18), are within a few percentage points error, up to about 20 carbon atoms for paraffins and 14 carbon atoms per molecular structure for all others. For determining PC and TC from a mixture, having a known PC and TC for each component, use molar percentages of each component times the respective PC and TC. Then add these PC and TC values to get the sum PC and sum TC of the mixture. Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 18

Chapter One

Molecular Weight The molecular weight of a discrete component or a group mixture is a very basic and indeed needed data input for defining any component. It is mandatorily defined in any characterization or assay-type hydrocarbon analysis. Molecular weight is indeed a must for solving any fluidtransport design problem. It, together with the subject fluid’s boiling point temperature, is the most important data to have or determine. I propose that molecular weight can be determined by means of two methods. Table 1.3 is again the first method proposed. If your component is not specifically listed in Table 1.3, simply estimate using the other similar family-type compounds to secure the MW. Referring to Table 1.3, please note that molecular weight values are 120 or less for all compounds. The API Technical Data Book lists many more HC compounds of 120 MW or less. Compounds of this type should receive MW determinations using these tables, referring to Table 1.3 and Ref. 4. The second method I propose to determine MW is the crude characterization method. For the past six decades, we have relied on the standard ASTM D86 distillation test to characterize crude petroleum and its products [6]. The next section includes excerpts from the ASTM4 program for crude oil characterization presented in the CD. Please note that there is a proposed MW equation on line 4690. I find this equation to be reasonably accurate, ⫾3% or less, for most every HC compound or HC pseudogroup above 120 MW. The ASTM4 printout in the next section, in Table 1.5, shows a run for a typical ASTM D86 lab analysis of crude oil. Use this program with caution, however, especially for compounds 100 MW or less. Errors here may exceed 10% in this region. The ASTM4 program is derived from Fig. 2B2.1 in Chap. 2 of the API Technical Data Book [15]. I have derived the equation in line 4690 using a curve-fit math which checks very well with the API figure. This API book is historically a good and reliable source for ASTM crude MW determination. Thus I have included this curve-fitted equation here in the ASTM4 program as the calculation for molecular weight. As seen in ASTM4 line 4690, the API gravity and the average boiling temperature are all that is needed. These two variables, gravity and boiling point, are commonly determined in every lab ASTM analysis run for any crude oil cut, hydrocarbon, or petroleum product. The equation I have presented checks very well within the range it was intended for, ASTM D86–type distillation cuts. One last MW note I wish to leave with the careful reader: For many years I have been asked to consult on difficult refinery problems concerning naphtha, gasoline, jet fuel, diesel, and gas oil petroleum cuts.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

19

In each and every one of these problems, I have requested immediate petroleum light ends chromatographic and ASTM D86 lab tests from the client’s lab services. In answer to half of these requests, I have been handed a document used for the design of the refinery that was at least 15 to 20 years out of date with current operations. While being handed these archaic wonders, I have been told, “We don’t get the ASTM test you requested from our lab, and they don’t get the proper samples to run the test you requested.” I have been even further moved by the fact that these laboratory marvels didn’t have the proper apparatus nor the personnel experience in their facilities to run such simple ASTM tests. Well, I must now say that the light ends, C1, C2, C3, C4, and so on, demand a gas sample–type chromatograph laboratory test. I strongly encourage that a well-experienced and reputable lab service company be contracted to run not only the light ends but also the full ASTM distillation test involving the heavier crude for C6 and the higher boiling point cuts. As a normal service, these professional labs include the cut gravity and the boiling points of each distilled cut logged in the ASTM test. Boiling Point The boiling point is the last data herein sought out; however, it is indeed the most important data to secure for a discrete pure component or a pseudo–crude cut component. Since the discrete pure components are generally a known type of molecular structure, their boiling points may readily be obtained or estimated from data sets such as Table 1.3. The crude oil components are left, unfortunately, undefined. Therefore, this section is dedicated to defining the boiling points of crude oil and its products. Over the past six decades, the petroleum industry has defined the many components making up crude oil and derived products simply by defining boiling range cuts. Each crude oil cut is generally held to a 20°F or less boiling range increment of the total crude oil sample or the crude oil product sample. Each of these boiling ranges is defined by a pseudocomponent. This is not a true single component, but rather a mixture of a type of components. Each grouped pseudocomponent mixture is treated throughout calculation evaluations as a single component. These pseudocomponents thus become the key database defining all petroleum processing equipment and processing technologies. There are two principal types of crude oil boiling point analysis. These are the American Standard Testing Method (ASTM) and true boiling point (TBP) test procedures [6]. First, the ASTM boiling analysis is actually two tests, the D86 and the D1160 [6]. The D86 is performed at atmospheric pressure, wherein the sample is simply boiled out of a container flask and totally con-

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 20

Chapter One

densed in a receiving flask. As the D86 involves cracking or molecular breakdown of the crude sample at temperatures above 500°F, this test method is extremely inaccurate at temperatures of 450°F and above. Thus, a second test method has been added to the D86, the D1160, which uses a vacuum for the same sample distillation at temperatures 450°F and above. The D1160 uses the same setup as the D86, only with an overhead vacuum, usually 40 mmHg absolute pressure. A good complete ASTM analysis of a normal virgin crude oil sample would thus involve starting with a D86 at atmospheric pressure and finishing with a vacuum D1160 when the D86 boiling temperature reaches 450°F. In the 1930s and earlier, the D86 ASTM–type test was discovered to be inaccurate regarding defining the true boiling ranges of these sodefined pseudocomponents. This inaccuracy is totally due to the fact that every pseudocomponent boiling mixture has boiling range components from its adjacent pseudo cuts. How can this test problem be solved? The answer is simple. Just run a TBP. How can this be done? Use a fractionation-type separation lab setup, refluxing the overhead boilout, which produces a more truly defined boiling cut range pseudocomponent. Thus, for each of the cuts, 20°F or less, a more accurate database of pseudocomponents is so defined. TBP-type lab tests are more the current-day standard of reputable labs. The Bureau of Mines of the U.S. government uses the Hempel TBP method. The ASTM commission adapted a method called the D2892. Both methods are similar, starting with overhead atmospheric pressure and finishing with vacuum, 40 mmHg or less. Both use trays or internal packing which is refluxed with overhead condensing. The fractionating column is maintained at a stabilized temperature as the temperature profiles of the column increase per distillation boilout progress. The outcome of this test is the true boiling point pseudocomponent definition. The D2892 lab test is a rather difficult test to run, requiring extensive laboratory work and considerable specialized equipment. It is therefore most apparent that the simpler ASTM tests D86 and D1160 are preferred. No fractionator reflux is required. These ASTM distillations, compared to the more rigorous TBP tests, are more widely used. This is because the ASTM tests are simpler, less expensive, require less sample, and require approximately one-tenth as much time. Also, these ASTM tests are standardized, whereas TBP distillations vary appreciably in procedure and apparatus. In earlier years, API set up calculation procedures to convert these more easily run ASTM D86 and D1160 boiling point curves to the sought TBP curve data. This book presents a unique program, named ASTM4, which receives ASTM curve inputs, both D86 and D1160 data, and converts them to TBP data. API 3A1.1 and API 3A2.1 methods are

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

21

referenced. Referencing ASTM4 starting at line 1660 (Table 1.6 in the next subsection), the ASTM 50% points are converted to the TBP 50% points. Then TBP point segment differences with the ASTM D86 and D1160 segment points are established. The end result is that the user may input ASTM D86 and 1160 data into the program and derive answers as though the input were TBP data. ASTM4 was written by the author and is based on derived mathematical algorithms which simulate the method in Chap. 3 of the API Technical Data Book [7]. The reader is referred to the API book for a more detailed derivation of the results from the API curves. Curvefitted equations simulating these API curves are installed in the ASTM4 program and give reasonably close or identical results. This was close to an exhaustive method of deriving a program for generating the needed crude oil pseudocomponent data. But, in due respect to all of those very finite, reliable, accurate, and costly simulation programs, ASTM4 does not equal their perfection. However, as the quest of this book is practical process engineering, ASTM4 produces reasonable and similar results as these high-end programs produce. I, the author, do bow to their excellence, but also imply that ASTM4 earns the right to be compared and in some cases may even produce equal results for the more complex and difficult crude oil runs. Please note we have been comparing and reviewing the API crude oil characterization method. There are several other crude characterization methods, including those of Edmister and Cavit. These however have all been compared, and the API data method upgraded ever closer to perfection as time has allowed. ASTM4.exe is a PC-format computer program for the ASTM D86 and D1160, or the TBP curve point input. The TBP method is that of the API group, having at least 15 theoretical stages and at least a 5-to-1 reflux ratio or greater. A typical example of ASTM4 input is given in Table 1.4, followed by a refinery gasoline stabilizer bottoms cut having a 10,000-barrel-per-day (bpd) flow rate in Table 1.5. Please note that there are nine points input. Each point is given an ASTM boiling point from a D86 curve and the volume percent that has boiled over into the condensate flask and the accumulative °API gravity reading of the boiled-over fluid. Please note that nine ASTM curve points are input. Also note that this ASTM D86 lab test is extended well over the 550°F temperature, which is the component cracking and degrading temperature. It would have been much more prudent to have stopped the atmospheric distillation at 550 and used a vacuum procedure such as the previously discussed D1160. A disclaimer, however, is made here for this being an actual lab test in which the lab indeed made a D1160 test for all components above the 500°F temperature and converted all the results to a simple nine-point single ASTM D86 test result as shown. The result

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 22

Chapter One

TABLE 1.4

ASTM4 Input

No.

°F

Vol, %

API

Dist, mmHg

1 2 3 4 5 6 7 8 9

290 380 450 510 575 640 710 760 840

10 20 30 40 50 60 70 80 99

50 40 36 32 30 29 24 23 18

760 760 760 760 760 760 760 760 760

NOTE:

ASTM curve points input = 9; mid-BP curve point = 5; bpd rate = 10,000.

TABLE 1.5

ASTM4 Output Answers

No.

mol/h

TC, °R

PC, psia

MW

Acent

API

Vol BP, °F

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

1.6400e+02 1.7352e+02 1.6921e+02 1.6496e+02 1.6077e+02 1.5663e+02 1.7803e+02 1.7171e+02 1.6557e+02 1.5962e+02 1.6776e+02 1.6186e+02 1.5618e+02 1.5074e+02 1.4554e+02 1.4203e+02 1.3640e+02 1.3102e+02 1.2588e+02 1.2028e+02 1.1513e+02 1.1023e+02 1.0555e+02 1.0143e+02 9.8444e+01 9.5679e+01 9.3139e+01 1.1644e+02 1.1212e+02 1.0806e+02 1.0424e+02 1.4121e+02 1.3862e+02 2.0792e+02

1019 1042 1066 1090 1114 1137 1159 1180 1200 1221 1241 1261 1281 1300 1319 1336 1353 1369 1385 1400 1415 1429 1442 1456 1472 1486 1501 1514 1524 1534 1544 1555 1565 1574

475 462 449 436 423 411 403 389 375 361 348 336 324 312 302 289 276 264 253 241 229 217 206 197 191 186 181 176 168 159 151 147 144 142

116.2 120.32 124.65 129.18 133.93 138.91 144.65 150.73 157.1 163.78 170.75 177.96 185.46 193.23 201.28 210.15 219.46 229.14 239.19 250.04 261.59 273.63 286.17 298.47 309.8 321.13 332.36 343.76 357.68 371.85 386.23 397.81 408.15 417.9

0.29779 0.31021 0.32179 0.33258 0.34262 0.35197 0.37321 0.38808 0.40321 0.41868 0.4345 0.4506 0.46729 0.48471 0.50298 0.5236 0.54539 0.5683 0.59248 0.61847 0.64604 0.67509 0.70577 0.73896 0.77647 0.81838 0.86551 0.9183 0.97295 1.03346 1.10091 1.18605 1.28789 1.40839

49.2 47.4 45.6 43.8 42.0 40.2 39.2 38.4 37.5 36.7 35.8 34.9 34.0 33.0 32.1 31.6 31.1 30.6 30.2 29.8 29.6 29.4 29.1 28.5 27.4 26.2 25.0 24.0 23.7 23.4 23.1 22.1 21.0 19.9

220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640 660 680 700 720 740 760 780 800 820 840 860 880

Totals: mol/h = 4.7499e+02; lb/h = 1.0974e+05; MW = 231.0

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database The Database

23

is therefore a smooth curve with nine points taken. The points above the cracking temperature, 500°F, are actually tested under a vacuum and converted to the shown ASTM D86 test. No cracking has occurred in this test. Aren’t professional labs wonderful? ASTM4 has supplied you with pseudocomponent characterization, molecular weights, acentric factors, critical constants, boiling points, and pseudocomponent gravity. With this database you are prepared to resolve most any crude oil and products database problem, deriving calculated needed results. Crude oil characterization— brief description of ASTM4

Some may ask why I’ve named this program ASTM4. It is indeed a computer computation method to define or simply characterize petroleum and its products. But why the 4? Because this is my fourth-generation upgrade of the program. The following will give the reader a brief ASTM4 program walkthrough: Making the distillation curve and API gravity curve. ASTM4 is set up for

an ASTM D86 distillation curve input. Although the TBP curve input could be set up as an option, this has not been done. The user could, however, make this option input if desired. The DOS version of ASTM4 does indeed offer this option of TBP input. I suggest that the user, if desired, may follow the same pattern as shown in the DOS ASTM4 version, adding a few option steps. The first lines of the program, to line number 2350, convert the ASTM data points to a TBP database. (See the program code listing in Table 1.6.) As seen in these line codes, curve-fitted equations are applied extensively. The bases used are API Technical Data Book Figs. 3A1.1, 3A2.1, 3B1.1, 3B1.2, 3B2.1, and 3B2.2. Code lines 2370 through 2400 correct the ASTM distillation for subatmospheric pressures. Instead of inputting 760 mmHg as shown in the example, you could also input 0 for the same results. The reason for a 0 input is line 2370, where this correction is bypassed. The reference for this subatmospheric pressure correction is the API Fig. 3B2.2 routine. The equations shown are curvefitted. In fact, this could be a D1160 distillation conversion to the TBP database. If the user desires to make an optional TBP direct database input in place of an ASTM database input, then skip from line 1650 to 2360. The DOS ASTM4 version in the CD disk offers this option. The next lines of the program, to line number 2480, establish curvefitted equations. Refer to code lines 2420 through 2490. (See the program code listing in Table 1.7.) Note that linear equations between each of the given ASTM data points are made, giving an overall ASTM

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com) Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

The Database 24

Chapter One

TABLE 1.6

ASTM4 Program Code, Lines 1040 to 2350

40 MCT = 0: WCT = 0 1040 For I = 1 To N2 1610 If B(I, 1) 20 GoTo 1850 1800 If B(I, 2)